Section 2 · Set up Database Observability

How Amazon RDS, Aurora, Cloud SQL, and Azure Database differ from the self-managed setup path

Estimated time: 2 min

Set up a managed database

The hands-on paths on the MySQL and PostgreSQL slides are for a database you run yourself. Amazon RDS, Amazon Aurora, Google Cloud SQL, and Azure Database use the same Database Observability features, and each has its own setup guide.

Each of those guides follows the same three steps:

  1. Prepare the database. Set the parameters or flags the guide lists, restart the instance when a change requires it, and create the monitoring user. For PostgreSQL, enable the pg_stat_statements extension in each database you want to monitor.
  2. Configure Grafana Alloy so it can reach that instance.
  3. Verify that query metrics appear in Database Observability.

Reach a Cloud SQL instance

Google Cloud SQL can be reached by private IP, by public IP with authorized networks, or by the Cloud SQL Auth Proxy. When you use the proxy, point Alloy at the proxy address. Avoid sending Alloy through a load balancer or a connection pooler.

Skip the provider’s system database

Grafana Alloy 1.17 and later excludes these provider system databases by default. If you run an older Alloy, exclude the database for your service yourself. The setup guide for your service shows where to set that exclusion.

ServiceDatabase to exclude
Amazon RDS and Amazon Aurorardsadmin
Google Cloud SQLcloudsqladmin
Azure Database (MySQL and PostgreSQL)azure_sys and azure_maintenance

Setup guides

Choose the guide that matches your engine and your service: